Materials development for Solid Oxide Fuel Cells -Status and development perspectives
Prof. Dr. Robert Steinberger-WilckensCentre for Hydrogen & Fuel Cell ResearchUniversity of Birmingham
Molecular Aspects of Solid State and Interfacial ElectrochemistryDubna, 26th – 31st August 2012
Slide 2/43Dubna 2012
Overview
• introduction to SOFC• fuel cell applications and their requirements• fuel cell problems and development goals • materials development for SOFC• understanding fuel cell degradation
Slide 3/43Dubna 2012
What is a ‘fuel cell’ and what does it do?
Slide 4/43Dubna 2012
Ele
ctro
lyte
porousanode
porouscathode
O2 (air)
Surplus air
H+
Fuel Cell Principle
+ H2O
membrane properties:• gas tight• high ionic
conductivity• low electronic
conductivity
H2
Surplus fuel
PEFC:typical ‚hydrogen fuel cell‘
‚membrane‘
Slide 5/43Dubna 2012
Ele
ctro
lyte
porousanode
porouscathode
H2
+
O2 (air)
Surplus air
O--
Ionic conductivity = f(T)
T = 400 ... 1000°C
Solid Oxide Fuel Cell
CH4, CO,
CO2, H2O
direction of current flow is identical to PEFC!
Surplus fuel
Slide 6/43Dubna 2012
Potential of SOFC in the Future Energy System
• fuel flexibility (H2, CH4, CnHm, CO, diesel, petrol ...)• minimal need for fuel processing for small residential CHP,
portable units, APU etc.• high electrical efficiency up to 60% (system, net)• role in transition strategies from fossil feedstock
to renewables and to hydrogen (including bio-fuels of various origin, liquid or gaseous, and hydrogen)
• fuel impurity tolerance• applications range from small scale residential CHP, APU and
portable (SOFC) to large units in industrial CHP and bulk power production (SOFC and MCFC)
Slide 7/43Dubna 2012
European SOFC Stack Technologies
AIRVariety of manufacturers and design types
• planar stacks - higher performance- compact design- mechanically robust- simple manifolding- lower cost
• tubular stacks- resistant to high temperature gradients (*)
- mechanically robust (*)- low power density
(*) thermo-mechanical stability greatly depends on SIZE, not so much on concept
Slide 8/43Dubna 2012
Variety of SOFC Cell Concepts
YSZ
LSM
Ni + YSZ
Electrolyte supported~ 300µm
ESC
Anode supported
600 µm – 1 mm
ASC
LSCF
YSZ/SSZNi + YSZ
CGO
Metal supported~ 1 mm
MSC
LSCF
YSZ/SSZNi + YSZ
FeCr
Specific properties with different application opportunities
Thin films on thin substrate
~ 300 µm
ASC
LSC/xSCF
SSZNi + SSZ
CGO
1000 °C 700 °CTemperature 700 °C 400 °C
CGO
barrier
Slide 9/43Dubna 2012
SOFC Applications
stationary
portable
Slide 10/43Dubna 2012
Performance of ‚Conventional‘ Products
Service life• vehicles >10 years (5.000 to 10.000 operating hrs)• heating boilers (residential power) >10 years
(20.000 to 40.000 hrs, frequent cycles possible)• power generating equipment 10 – 30 years (40.000
to 200.000 operating hours)
Other• vibration and shock (road vehicles)• acceleration (aircraft)• simple coupling to natural gas supply
(boilers/engines)
Slide 11/43Dubna 2012
SOFC Development Challenges
• improved durability under static, transient and cycling conditions- redox stability- thermal cycling capability
• stack lifetime in excess of 40.000 hrs. (stationary & large units, loss of power at end of life <20%)
• high performance, high efficiency• arbitrary switch-off and start-up cycles (several 100 to 1000)• tolerance against fuel impurities• operation without external water supply • robustness to vibration and mechanical shock• design of large units and hybrid power plants• lower cost, increased system compactness, simplification of
technology
topics in joint materials,
design and systems
development
Slide 12/43Dubna 2012
HEXIS: Comparison of ZIP Stack Generations (2000/2002)
Slide 13/43Dubna 2012
Real-SOFC Stack Generations: Progress by Materials
700°CLSCF
Slide 14/43Dubna 2012
Materials development for SOFC
Slide 15/43Dubna 2012
Increased Performance through Improved Materials
• low ASR through* low cathode overpotential -> high oxygen ion transfer rates* high conductivity of electrodes* thin layers
• electrolytes with higher conductivity• hermetic separation of layers
-> thinner layers of highly active but reacting materials-> interdiffusion barriers
• mechanically stable contact layers with high electric conductivity• higher performance at lower temperatures -> less degradatoin
Slide 16/43Dubna 2012
Materials for Increases in Performance
600 650 700 750 800 850 900650 700 750 800 850 9000.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
Mean Value 2005 - 2008
Curr
ent D
ensi
ty70
0 m
V / A
/cm
²
Temperature / °C
series 01 series 02 series 03 series 04 series 05
SOFC with LSCF cathode
LSC (2009)
from LSM to LSC:Lanthanum-Strontium-Manganite .... Lanthanum-Strontium-Cobaltite
Slide 17/43Dubna 2012
Materials Processing: Diffusion Barrier for LSFC Cathodes
YSZCGO
EB-PVD of CGO layer at target temperature 800°C
1 m
1 m
CGO, 800 °C
YSZ
EB - PVD layers: thin, dense, gas tight structure, strong bonding of YSZ & CGO layer
Slide 18/43Dubna 2012
Thermal Cycling Requirements • thermal cycles:
- ‘cold start’ 20°C … 200°C, ‘warm start’ >400°C up to 600 … 750°C• goals:
- no gas leakages from stack (safe operation)- rapid start-up (30 minutes for road APU)
768air_in
air_in
air_out
fuel_out
fuel_in
fuel_in
measured temperatures
minimal calculated temperature
808.5
FEM-simulation of this part
795°C
highest tensile stress
leakages
solutions:• strong sealings• robust design• compliant design
achievement:• 100 to 250 cycles > 25°C(JÜLICH, ElringKlinger)
Slide 19/43Dubna 2012
Strengthened Glasses
1. Fiber reinforcement of Ba-Ca-Silicate glass matrix by YSZ fibers
• Reduced crystallization kineticsof matrix
• Low porosity of the joint• Minimal interactions of fibersLinear correlation between thermal expansion and amount of filler
2. Doping of glass with ductile material (e.g. silver)
• increased strength• but also increased conductivity
YSZ fibers
Residual Glass Phase
Slide 20/43Dubna 2012
Redox Cycling Requirements • redox cycles:
- after stack shut-down air will flow to the fuel electrode- Ni in Ni-YSZ anode will re-oxidise to NiO2- NiO2 has higher volume and will cause mechanical damage to cell
• goals: - no gas leakages from stack (safe operation)- rapid start-up (30 minutes for road APU)
Elektrolyte
Anode
Substrate solutions:• system control of
temperature and fuel flow• robust cells
Slide 21/43Dubna 2012
Anode Redox Stability – SrTi Anode
-1500
-1000
-500
0
500
1000
1500
455 465 475 485
Time in h
Vol
tage
in m
V
iV
Full redox cycles cell_voltageO2_inO2_out
Anode Redox Stability – SrTi Anode
disadvantage:low electrochemical performance due to low electrical conductivity
Slide 22/43Dubna 2012
Consolation of Conflicting Properties
low degradation high performance
good handling and processing propertiesimproved
robustness(cost)
for instance: redox stable materials (SrTi, LSMC),
with low conductivity and brittle structure
Slide 23/43Dubna 2012
Interaction of Materials Developers and Manufacturers
‚improved‘ or ‚promising‘ material
processing properties
modified ‚compromise‘ material
processing ‚tricks‘
composite materials
functional ‚improved‘ component
building a bridge from materials research to component manufacturing
Slide 24/43Dubna 2012
Outlook
Materialscurrently best performing materials have already been known for many years (no surprises)optimisation is necessary with respect to processing and costLifetime is still insufficient (but: trade-off with cost)breakthroughs are nevertheless necessary (new materials integrated with processing and manufacturing)
RTD challengespurpose-designed materials incl. ab-initio understandinglow-cost, standardised, mass-production oriented manufacturing extended lifetime of components, robustnesssufficient testing capacity for reliably & rapidly predicting materials performance (optimisation loops!)
Slide 25/43Dubna 2012
Project N-KATH
cooperation between FZ-Juelich, MSU and BIC, and company HC Starck‘design’ of cathode perovskite material according to theoretical considerations and modelssynthesis of materialsverification in SOFC cell experiments
La2CuO4 Pr2CuO4Pr1.6Sr0.4CuO4
Layered perovskites: which structure blocks are necessary for good O-conductivity?
Slide 26/43Dubna 2012
Understanding fuel cell degradation
Slide 27/43Dubna 2012
Stack repeating unit
SOFC repeating unit components to be addressed and details of the specific layers that interface with each other
Slide 28/43Dubna 2012
Variety of Degradation Phenomena
50 100 150 200 250 300 350 4000.0
0.2
0.4
0.6
0.8
1.0
1.2
0
2
4
6
8
10
Voltage
Volta
ge
Time (Hours)
Current
sulphur poisoning
chromiumpoisoning
insufficient contacting
anode re-oxidation
corrosion
Slide 29/43Dubna 2012
The quantification and prediction of single contributions with respect to their behaviour over time is the key expected outcome of this project
Single effect experimental isolation,
sensitivity matrix
Description of changes in properties
X = f ( t, T, i, p(O2), uF, …)Y = f ( t, T, i, p(O2), uF, …)
Electrochemical model
EMF = F (X, Y, t, T, …)
Prediction&
Correlation
Perf
orm
ance
Time
segmented cells
Choice of testing conditionsCorrelation
Single effect experimental isolation,
sensitivity matrix
Description of changes in properties
X = f ( t, T, i, p(O2), uF, …)Y = f ( t, T, i, p(O2), uF, …)
Electrochemical model
EMF = F (X, Y, t, T, …)
Prediction&
Correlation
Perf
orm
ance
Time
segmented cells
Choice of testing conditionsCorrelation
Understanding degradation
Slide 30/43Dubna 2012
Degradation types
1. continuous, steady degradation- initialisation phase (sintering, saturation)- constant slope phase- progressive degradation phase (EoL)
2. degradation after ‚events‘- thermal cycle- redox cycle
3. degradation after ‚incidents‘- malfunction of BoP components- malfunction of control- external influence (shock, grid outage etc.)
Slide 31/43Dubna 2012
Cathode Materials: Stability
(La,Sr)MnO3
(La,Sr)FeO3
(La,Sr)CoO3
(La,Sr)(Co,Fe)O3
source: Yokokawa, EMPA
thermodynamical stability and kinetics:perovskites ABO3
(La0.9,Sr0.1)MnO3
(La0.7,Sr0.3)MnO3
(La0.7,Sr0.3)0.99MnO3
(La0.7,Sr0.3)MnO3±δ
(La1-xSrx)yFe1-z(Ni,Cu)zO3-δ
Slide 32/43Dubna 2012
Cathode Materials: Volatility
source: Tietz/Mai
gas flow
Sr deposition
Slide 33/43Dubna 2012
K.S. Lee et al. / J. Solid State Electrochem. 11 (2007)1295
Anode Substrate: Particle Agglomeration
• temperature-induced tendency of metals to decrease free energy, i.e. to minimize the surface area and agglomerate
• examples: anode substrate Ni-YSZ cermet
10 µm
heated up for 4000 hat 1000°C in
Ar/4%H2/4%H2O
Ni: whiteYSZ: greypores: black
new cermet
Slide 34/43Dubna 2012
Three-Dimensional Characterisation
J.R. Wilson et al. / Nature Materials 5(2006)541
• FIB/TEM analysis• reconstruction of
3-D structure from ‚slices‘
Slide 35/43Dubna 2012
LSM/YSZ
(CrMn)3O4 (spinel)
electrolyte (YSZ)
Chromium Poisoning: Microscopic Findings
Slide 36/43Dubna 2012
Sulphur Poisoning – The Phenomenon
50 100 150 200 250 300 350 4000.0
0.2
0.4
0.6
0.8
1.0
1.2
0
2
4
6
8
10
Voltage
Volta
ge
Time (Hours)
Current
Slide 37/43Dubna 2012
Sulphur Poisoning: Microscopic Findings
bulk materialdeposition► Reduction of porosity► Elimination of
catalytically active Ni
surfacedeposition► Elimination of
catalytically active Ni
Slide 38/43Dubna 2012
Coking – Carbon Buildup in Internal Reformingcarbon build-up due to hydrogen and oxygen stochiometry mismatch (Boudouard Reaction)
figures courtesy of Jörger & He
Slide 39/43Dubna 2012
Cyclic Oxidation of Ferritic Steel Crofer 22 APU in Air at 900°C
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
0 200 400 600 800 1000 1200 1400
Time (h)
Wei
ght c
hang
e (m
g/cm
2 )
0.3 mm(breakaway)
0.5 mm
2 mm
0.1 mm(breakaway)
kp–dependence on specimen thickness
Break-Away Corrosion
Slide 40/43Dubna 2012
grain boundary oxidation front
etched sampleJS-3
CroFer22APU 1st
glass remains
150hH2/H2O
optimal matching of steel and sealing materials isvital:- good adhesion = chemical interaction- but: no excessive corrosion
Interaction of Glass Sealant and Ferritic Interconnect
Slide 41/43Dubna 2012
High Degradation due to Contacting problemsContact trace on cathode
2 mm
880 mV - 298 mA/cm²770 mV - 336 mA/cm²820 mV - 336 mA/cm²
high local current due to narrow contacting ‚ridge‘
Slide 42/43Dubna 2012
cut
air intake air outlet
G`1002-03Thermo-Mechanics
low strength of steels at high temperatures
Slide 43/43Dubna 2012
Conclusions
materials development is crucial in improving the performance of electrochemical devices (like fuel cells)developments have to be coordinated with practical aspects of technologythe understanding of materials behaviour is just as important as the development of ‘new’ materialsmicroscopy and tomography are essential tools in doing solifetime modelling can help in developing accelerated testing and prediction methods for materials and components
Slide 44/43Dubna 2012
Acknowledgments go to my former project team at FZJ Dmitri Bronin for slide material
to the European Unionfor financing part of the research presented here,
and to all partners in the co-operation projects Real-SOFC, SOFC600, SOFC-Life, ACCELENT,
MMLCR=SOFC, N-KATH
Real-SOFC was co-financed by the European Commission under the contract no. SES6-CT-2003-502612
SOFC-Life is co-financed by the Fuel Cell and Hydrogen Joint Undertaking (FCH JU) under the contract no. 256694
Thanks for your Attention!
